« 1 quick but powerful test design tip | Home | 3 difficult optimization results and what you can learn from them (1 of 3) »
3 parts to picking a test page
By Billy | March 26, 2008
Alright, so you’re ready to test. You’ve got tools and the skills to design a test. But when optimization begins, where should you start? While we all would love ROI to be the only driving factor in optimization, your resources and reach usually dictate what you end up testing, as well as ROI.
Here’s what I think about when looking for candidate test pages:
- Importance: Is this the best page to accomplish your goals? Take a look at your overall marketing campaign and see how important this is to the whole process. Look at drop-off points and find pages that are important but weak links.
- Technical: Will this page be easy to optimize? How much technical involvement will it require? If you’re testing dynamic elements you may need some additional help. Maybe you can test a page outside of the development schedule or on a separate server. Look for pages that have less restrictions and can be modified quickly. Also examine the page for what can’t be tested and what you may want to test on a page. Some tests are harder to create than others, both technically and creatively.
- Goals: What are you optimizing for? Pages with one goal are easier to optimize since you can drive everything on the page towards achieving that one goal. If you think a page is under performing, then it may be an easier page to optimize also. Lastly, think about how easy that goal will be to measure. If there are multiple conversion possibilities or the conversions are offline, it will be difficult to test.
In the end you are asking a multi-part question: Will a lift here be more valuable than a smaller/same/larger lift elsewhere that will take X amount of work and time?
Just remember, you can always test a page later on, even if it may not be the best candidate now. If you have the resources you can test pages simultaneously too. Just make sure they don’t impact each other in any way, so as to not skew your results.
As your optimize more and more, it will be harder to choose, but that’s a good sign. It means you’ve got a lot of great pages and that is what you want testing to do for you.
Related posts
Topics: Methodology |